LOWER CRETACEOUS SPIDERS FROM THE SIERRA DE MONTSECH, NORTH-EAST SPAIN

Transcription

1 LOWER CRETACEOUS SPIDERS FROM THE SIERRA DE MONTSECH, NORTH-EAST SPAIN by PAUL A. SELDEN ABSTRACT. Four new specimens of spiders (Chelicerata: Araneae), from Lower Cretaceous (Berriasian-Valanginian) lithographic limestones of the Sierra de Montsech, Lérida Province, north-east Spain, are described, as Cretaraneus vilaltae gen. et sp. nov., Macryphantes cowdeni gen. et sp. nov. (two specimens), and Palaeouloborus lacasae gen. et sp. nov. All belong to the infraorder Araneomorphae. Palaeouloborus is the oldest representative of the superfamily Deinopoidea, Cretaraneus is referred to the superfamily Araneoidea, and Macryphantes is the oldest record of the superfamily Araneoidea, family Tetragnathidae (metine-tetragnathine-nephiline group). All three spiders were web weavers; Macryphantes and Palaeouloborus wove orb webs, and may have used a wrap attack to prey on the abundant contemporaneous insect life preserved in the Montsech deposit. SPIDERS are rare in rocks of Mesozoic age. Only two specimens, Juraraneus rasnitsyni Eskov, 1984, and Juratchaea zherikhini Eskov, 1987, from the Jurassic of the USSR, have been adequately described. Three specimens from the Lower Cretaceous of New South Wales (Jell and Duncan 1986) and twelve from the Trias of France (Gall 1971) were not identified to a taxonomic rank below that of order. Spiders reported from Canadian Cretaceous amber (McAlpine and Martin 1969), the Jurassic and Cretaceous of Siberia and Mongolia (reported in Eskov 1984), and the Cretaceous of Botswana (Waters 1989) await description. The four specimens described here are sufficiently well preserved for their taxonomic affinities to be determined with some precision, and thus they represent only the third find of Mesozoic spiders to be described and named. The fossil spiders described here are placed in extant superfamilies or families, but closer assignation is considered unwise pending the outcome of current studies on living and fossil members of these groups. GEOLOGICAL SETTING Locality. The Sierra de Montsech lies in the foothills of the eastern Pyrenees, between Balaguer and Tremp in Lérida Province, north-east Spain (see Schairer and Janicke (1970) for details and location map). Three quarries in the vicinity of the abandoned village of Rubies, in the eastern part of the sierra, yield remarkable fossils. 'La Pedrera de Meiá' and 'La Cabrúa' have been worked for many years, and 'El Reguer' is currently under investigation. Stratigraphy. The 50 m succession of limestones exposed in the quarries, the ' Calcaires lithographiques á Plantes et Vertebres de la Pedrera de Rúbies', is a facies development of the 100 m 'Calcaires á Charophytes du Montsech' (Peybernes 1976). The deposit has been determined as late Berriasian to early Valanginian in age on the evidence of ostracodes (Peybernes and Oertli 1972; Brenner, Goldmacher and Schroeder 1974; Whalley and Jarzembowski 1985), a conclusion consistent with evidence from palynology (Barale et al. 1984; R. Porter, personal communication 1988). Sedimentology. The Calcaires á Charophytes du Montsech is a sequence of mostly pale, finegrained, thinly-bedded limestones. Sedimentary structures include laminations, fine grading, minor deformation, and roll marks; trace fossils include arthropod trackways (Schairer and Janicke 1970). The environment of deposition deduced from the sedimentological and palaeontological evidence IPalaeontology, Vol. 33, Part 2, 1990, pp , 4 pis.] The Palaeontological Association

2 258 PALAEONTOLOGY, VOLUME 33 (see below) is that of coastal lagoons within a large area of algal flats separating marine conditions to the north from the Ebro continent to the south (Barale et al. 1984). The spider-bearing sequence, the Lithographic Limestones, represents a particularly tranquil lacustrine depositional episode (Lacasa and Martinez 1986). History of the fauna and flora. The exceptionally preserved biota of the Montsech lithographic limestone has received attention from palaeontologists since the beginning of this century. In the last few years, renewed interest in the deposit has brought new material to light, including the specimens described here, and avian fossils of phylogenetic importance from a locality of similar age and lithology in the neighbouring province of Cuenca (Sanz et al. 1988). Lists of the fauna and flora were given by Barale et al. (1984), with the most recent summary, especially of the insects, provided by Lacasa and Martinez (1986). All the indigenous microflora is of terrestrial or non-marine aquatic provenance; the macroflora includes a wide range of gymnosperms, progymnosperms, a few ferns and horsetails, and other, unclassified, plants. Animals include ostracodes, few decapod crustaceans, larval and mature insects belonging to eight extant orders, a wide variety of fish, a few frogs and reptiles, and some bird remains. The spiders were first mentioned by Lacasa (1985, p. 228), and a preliminary report of the results presented here was given by Selden (1989). MATERIAL AND METHODS Preservation. The spiders are preserved on thin slabs of pale buff-grey limestone. Grains are not Visible in the rock, and the hackly fracture and vitreous appearance under high-power microscopy suggest crystallization from a lime mud. Calcite-filled cracks cross some specimens. The spiders are preserved as pieces of cuticle on the bedding surface. The cuticle is brittle and brown: thicker parts are deep brown and the thinnest cuticle pale buff. The cuticle has not been chemically analysed; it is presumed still to be organic, but probably not of the original composition. The best-preserved parts are visible through a thin layer of translucent limestone, but their morphological details are hazy due to the presence of the overlying matrix. In such instances, 2-4 % hydrochloric acid was used, sparingly and with care, to remove the matrix and thus to reveal fine structural details. The spiders are in varying states of original decay: for example in LC 1753 AP the podomeres, of leg 1 in particular, are crowded with subcircular objects along the central parts of the shaft. These objects are interpreted as the decayed remnants of muscles. LC 1754 AP is very poorly preserved: mainly a mould remains, and this is interpreted as a specimen in which decay has progressed further. Both part and counterpart of specimens LC 1753 AP and LC 1754 AP are preserved, but only the part of specimens LC 1150 IEI and LP 1755 AP was collected. Specimens LC 1753 AP B, LC 1754 AP A, LC 1150 IEI, and LP 1755 AP represent lower slabs preserving mainly ventral features of the specimens; LC 1753 AP A and LC 1754 AP B are upper slabs with dorsal features. However, splitting of the rock has not resulted in perfect separation of dorsal and ventral, and due to the mode of preservation within the limestone (described above), most of both dorsal and ventral parts are preserved on LC 1753 AP B, and LC 1150 IEI. LC 1754 AP is mainly an external mould with little cuticle remaining, but on LP 1755 AP ventral parts of the body and both dorsal and ventral sides of the distal parts of the legs are preserved. On all specimens, the legs and abdomen (when present) are crushed flat. The carapace has sufficient convexity to produce relief in the fossils, so that on LP 1755 AP, for example, the carapace shape can be determined by the relief of the fossil, and the shapes of the sternum and coxae are outlined by setae and cuticle. The male palps of LC 1753 AP, LC 1754 AP, and LC 1150 IEI appear to occupy depressions in the matrix. This is probably because they were bulbous structures in life and therefore created a concavity in the sediment into which the palp collapsed during burial. The spiders were studied under a Wild M7S stereomicroscope, with the specimens immersed in ethanol or glycerine to enhance their contrast against the pale rock background. Camera lucida drawings were made and photographs were taken under the same conditions. In addition, a Zeiss photautomat microscope was used, in reflected light mode with oil immersion objectives, to view

3 SELDEN: SPANISH CRETACEOUS SPIDERS 259 and photograph details at higher magnifications. The computer program MacClade, version 2.1 (Maddison and Maddison 1987) proved very useful for exploring relationships. Terminology. Setal terminology is somewhat problematical, since different authors have used the same terms in different ways. Furthermore, there is a complete gradation in size between setae (small, and may be short or long, thick or thin), bristles, and spines (large). Macroseta is a term used, for example by Opell (1979), to describe a large seta which could equally be called a small spine or bristle. The common hair-covering of Cretaraneus is the serrate seta, which is smooth apart from a few minute accessory spines which are no longer than the mean thickness of the seta (see Lehtinen 1967, fig. 10a; Kullmann 1972, fig. 7). Two fairly distinct setal types are commonly called plumose (e.g. by Forster and Wilton 1973; Forster and Blest 1979). The first, which is here called plumose, is generally thicker than the serrate seta, and has helical ridges bearing small accessory spines which are much shorter than the width of the seta (see Lehtinen 1967, fig. 8; Kullmann 1972, fig. 8). The second, which is here called feathery, is fine, smooth, and has long accessory branches which are much longer than the width of the seta (see Lehtinen 1967, fig. 9). Many other types of seta and spine exist; there may be complete gradations between them, and the extent of their phylogenetic importance is unknown. Terminology of the sclerites of male palps differs between workers in different groups of spiders due to a lack of understanding of the homologies between the sclerites. Thus the task of recognizing palpal sclerites in fossil spiders is problematical. Useful descriptions of the constituent parts of male palpal organs are found in Comstock (1948), Levi (1961), Merrett (1963), and Millidge (1977). In leg formulae (e.g. 1243), the leg lengths are ranked in order longest (first) to shortest (last). Abbreviations used in the text and text-figures are as follows: ab abdomen 1 labium s serrula bo book-lung operculum lp left palp st sternum ca cephalic area m maxilla t tegulum cal calamistrum ma median apophysis ta tarsus eh chelicera mt metatarsus ti tibia ex coxa pa patella ti a tibial apophysis e embolus pc paracymbium tr trochanter f fovea pe pedicel fe femur rp right palp Provenance and depository. Three of the fossil specimens, LC 1150 IEI, LC 1753 AP, and LC 1754 AP, come from the quarry of La Cabrúa, the fourth, LP 1755 AP, is from the locality of La Pedrera de Meiá. Exact stratigraphical provenance is not known, but both of these localities are in the same 50 m sequence of lithographic limestones, the Calcaires lithographiques a Plantes et Vertebres de la Pedrera de Rúbies, described above. The specimens are deposited in the Institut d'estudis Ilerdencs, Lérida. Preserved specimens of extant spiders were studied for comparative purposes, and in addition to those in the author's collection of mainly British species, the following specimens were examined. Uloboridae: Hyptiotes ftavidus, female, Funchal, Madeira, M.J.Jones Collection No. 119, Manchester Museum; Uloborus walckenaerius, male and female, Chobham, Surrey, D. W. Mackie Collection No. G4999, Manchester Museum; Philoponella sp., male and female, Lake Naivasha, mature and immature males, Nairobi, Kenya, J. Murphy Collection Nos 1302, Deinopidae: Deinopis guianensis, female, British Guiana, British Museum (Natural History) (BM(NH)) No ; Deinopis stauntoni, female, Durban, South Africa, BM(NH) No ; Deinopis sp., female and immature, Kilifi, Kenya, J. Murphy Collection; Menneus camelus, females (types), Durban, South Africa, BM(NH) No ; Menneus sp., male, Kitale, and female, Nairobi, Kenya, J. Murphy Collection; Avella angulata, female, Gayndah, Australia, BM(NH) No ; Avella despiciens, female, Sydney, Australia, BM(NH) No

4 260 PALAEONTOLOGY, VOLUME 33 Dictynidae : Aebutina binotata, immatures, Aguas Negras, near Tarapuy, Napo, Ecuador, British Museum (Natural History) Arachnid Collection. SYSTEMATIC PALAEONTOLOGY Order ARANEAE Clerck, 1757 Suborder opisthothelae Pocock, 1892 Infraorder ARANEOMORPHAE Smith, 1902 Remarks. The phylogenetic scheme for the higher classification of spiders which is currently widely accepted originated with Pocock (1892). In this scheme, the Liphistiomorphae (Mesothelae) is the sister group to the Opisthothelae (Mygalomorphae + Araneomorphae). The morphology of liphistiomorphs is distinctive (Platnick and Gertsch 1976, fig. 5). The fossil spiders described here differ markedly from this pattern, lacking one or more of the following liphistiomorph synapomorphies: posteromedian invaginations of the fourth coxae, seven or eight forwardly situated spinnerets, sclerite of the first abdominal segment (Platnick and Gertsch 1976), and labium as wide as or wider than the sternum (Raven 1985). Thus they are referred to the Opisthothelae. Mygalomorph synapomorphies include the following external morphological features: cheliceral fang which operates in the sagittal plane (orthognathous), complete loss of expression of the anterior median spinnerets, three or four articles in the posterior lateral spinnerets (Platnick and Gertsch 1976), labial and maxillary cuspules, and reduced male palpal sclerites (Raven 1985). The alternative character states typify the Araneomorphae, and it is to this latter group that the Montsech spiders are referred. In particular, the chelicerae of Cretaraneus gen. nov. are labidognathous (i.e. they operate transversely); the male palp of Macryphantes gen. nov. is more complex than the mygalomorph palp; and Palaeouloborus gen. nov. possesses a calamistrum and a cribellum; the latter is a homologue of the anterior median spinnerets. Neither labial or maxillary cuspules nor three- or four-segmented posterior lateral spinnerets are known in any of the Montsech spiders, and each of them shows greater affinity to araneomorph families than to mygalomorphs. Assignment of the Cretaceous spiders to lower taxonomic categories presents greater difficulty. A competent araneologist can place most living spiders seen in the field into a family with a high degree of certainty. Even in the absence of field characters, such as web type and habitat, spider familial keys are relatively straightforward, only requiring the use of a binocular microscope and normally no dissection. Spider families are diagnosed on unique combinations of morphological characters such as number and position of eyes, spinneret pattern, and number of tarsal claws. However, research has shown that many of the characters on which the families are based are plesiomorphic at higher taxonomic levels, convergent with other groups, or are 'loss' characters (Lehtinen 1978). A classic example of this is the recognition of the cribellum as a synapomorphy for Araneomorphae: its presence in lower categories is merely the expression of the plesiomorphic state, and its absence illustrates convergence of the apomorphy in many araneomorph families. Whilst unique combinations of plesiomorphies may work as familial descriptions in practice, they cannot reflect true relationships between families. Of necessity, therefore, recent cladistic analyses have sought new or previously overlooked characters as apomorphies. These characters are commonly behavioural or anatomical, and may be poorly known in groups outside those under particular study. More important in the present work, such characters are most unlikely to be preserved in the fossil record. An additional concern encountered when working with rare fossils which have Recent relatives arises from the possibility that the fossils may be ancestors of Recent species. Should the fossils be classified in the same scheme as their Recent relatives, or in a separate scheme? Such problems have been discussed by Crowson (1970), Farris (1976), Wiley (1981) and others. A useful device for classifying fossils with their Recent relatives is the plesion (Patterson and Rosen 1977); this presupposes, however, that the Recent classification is soundly based on shared derived characters.

5 SELDEN: SPANISH CRETACEOUS SPIDERS 261 In spider systematica this is not yet true. In this paper, the new taxa are classified as close as possible to their presumed position in the Recent classification. Inevitably this entails placement within taxa yet to be defined by synapomorphies or defined on shared derived characters which cannot be seen easily in fossils. Because these are among the first Mesozoic spiders to be described, and because of the paucity of diagnostic features in the fossils, and the present lack of knowledge of synapomorphies in living families, the fossil spiders are assigned to superfamilies, and in one case to a family, but to no extant lower taxa. (Note that among Cretaceous insects, classification to modern genera is not unusual.) More specimens of Mesozoic spiders which are becoming available for study may help to elucidate further the taxonomic positions of the specimens described here, and assist in unravelling the complexities of spider phylogeny and evolution. The 'consensus' spider classification scheme given in Shear (1986Z>) is followed here in general, but with discussion concerning the status of groups assigned to the Araneidae following recent work by Coddington (1989a, b, 1990). Superfamily DEINOPOIDEA Koch, 1851 Remarks. This superfamily consists of the two cribellate families Uloboridae and Deinopidae, which weave orb webs (most Uloboridae) or spin modified orb webs which are then thrown, retiarius-like, at their prey (Deinopidae). Much more is known about uloborids than deinopids, though work on the latter is currently in progress. Although these two families have been considered closely related for nearly a century (Simon 1892), Coddington (1986, p. 359), with reference to his cladogram (p. 358), remarked that: 'the monophyly of the uloborid-deinopid lineage is based on only three characters, primarily because so little is known about deinopids.' The three characters Coddington accepted as deinopoid synapomorphies are: puffed cribellate silk, pseudoserrate plumose hairs, and fourth-tarsal macrosetae; however, he qualified this by suggesting that the first two characters may actually be synapomorphies for all orb-web weavers (Deinopoidea and Araneoidea) (Coddington 1986, pp. 327, 359). Recent work by Coddington (1989a, b, 1990) has resolved the apparent trichotomy between Araneoidea, Uloboridae, and Deinopidae (reported in Platnick 1986); Deinopoidea and Araneoidea being seen as sister groups in a monophyletic group of orb-web weavers: the Orbiculariae Walckenaer, Palaeouloborus gen. nov., described below, possesses three tarsal claws with accessory claws typical of web weavers and lacks the tarsal adaptations of the superfamilies Thomisoidea, Philodromoidea, Lycosoidea, Clubionoidea, Salticoidea, and Pholcoidea. The fossil genus also lacks the synapomorphies of Scytodoidea (Lehtinen 1986), Dysderoidea (Forster and Platnick 1985), Palpimanoidea (Forster and Platnick 1984), Hersilioidea, Hypochiloidea (Platnick 1977), Eresoidea, and Agelenoidea. Dictynoidea are cribellates, and the deinopoid families have, at one time or another, been referred to this superfamily, and to the family Dictynidae in particular. However, dictynids are generally small spiders, with short legs of approximately equal length, and lack femoral trichobothria (see below); thus they are quite unlike Palaeouloborus. The characters which are most useful in placing Palaeouloborus are: femoral trichobothria on legs 2, 3, and 4, leg 1 more than five times the length of the carapace and more than twice the length of leg 3, all leg tarsi with large accessory claws and apparently non-pectinate paired claws, presence of calamistrum and cribellum, and presence of plumose hairs. Femoral trichobothria occur in only two groups of araneomorph spiders: the metine-tetragnathine lineage of the superfamily Araneoidea, and the family Uloboridae (Opell 1979; Lehtinen 1980; Levi 1980, 1981). In both of these groups there is great disparity in length between the elongate anterior legs (1 and 2) and the short third pair (PI. 2, fig. 9), a feature also found in some other Araneoidea (Argiopinae, Levi 1983). Palaeouloborus is cribellate, and since only the Deinopoidea, but not the Araneoidea, are cribellate, this genus must be referred to the former superfamily. The calamistrum of Palaeouloborus is situated in a curved depression on the superior surface of the fourth metatarsus. Such a curvature occurs in uloborids (PI. 2, fig. 10), to a much lesser degree in deinopids (Shear 1986a), and also in Aebutina, a poorly known genus tentatively referred to the Dictynidae (only females and immatures of this genus are known). Additionally,

6 262 PALAEONTOLOGY, VOLUME 33 Palaeouloborus bears plumose setae, which are found in Deinopoidea but,not Araneoidea. The characteristic tarsal macrosetae of deinopoids (see below) appear to be absent from Palaeouloborus. Palaeouloborus is referred to the superfamily Deinopoidea, on the evidence given above. The fossil clearly resembles members of the family Uloboridae more closely than the Deinopidae; the latter family has many autapomorphies (e.g. forwardly directed, enlarged, posterior median eyes, elongate legs all of a similar length, web-throwing) and lacks the femoral trichobothria typical of the Uloboridae and the fossil genus. The position of Palaeouloborus within the Deinopoidea is now discussed. In a recent revision of the Uloboridae, Opell (1979) gave the following characters as diagnostic of the family: (1) lack of poison glands, (2) cribellate orb-web weavers, (3) femoral trichobothria, and (4) row of macrosetae (short spines) on metatarsus and tarsus of leg 4. The tarsal spines cannot be a synapomorphy for the Uloboridae since they also occur in Deinopidae (see below). Coddington (1986) added characters of the silk-glands described by Kovoor (1977), and some behavioural traits, to the list of uloborid synapomorphies, but omitted femoral trichobothria. Coddington's cladograms (1986, p. 358; 19896, fig. 108) show the metine-tetragnathines, which also have femoral trichobothria, far from the dichotomy of Araneoidea with Deinopoidea, which presumes that femoral trichobothria are a convergent phenomenon in uloborids and metines-tetragnathines. The alternative hypothesis (that they are a synapomorphy for all orb-web weavers) would require their loss in many separate lines. The row of tarsal and metatarsal macrosetae of uloborids was considered to be a synapomorphy of the family by Opell (1979), but in a later paper, Opell (1982) mentioned finding a poorly developed row on leg 4 of deinopids as well. My own observations confirm that macrosetae are present on the inferior surfaces of the distal half of the metatarsus and the tarsus of leg 4, and to a lesser extent on leg 3, in the deinopids Deinopis, Menneus, and Avella. These macrosetae are rather similar to the comb of serrate bristles seen in the Theridiidae in a similar position on the legs. However, they differ from theridiid bristles in being plumose, not serrate. As in the theridiids, they may need to be searched for, since they blend into the general hirsuteness of the tarsus. The macrosetae are not greatly different from the curved bristles normally present on the inferior surface of the distal end of the tarsus. Also, as in theridiids, they are not strictly confined to mt4 and ta4, also being present on leg 3, and they vary from species to species. In all the deinopids I studied they were quite unlike the comb of short, upstanding spines of the uloborid Zosis geniculatus illustrated by Opell (1979, pi. 1, figs. A and C). Rather, they resemble the row of macrosetae of the uloborid Hyptiotes cavatus figured by Opell (1982, pi. 1, fig. C). The similarity between theridiids and deinopids in this feature is presumably due to convergence in their prey-wrapping strategies rather than synapomorphy. Clearly, a comb of macrosetae on the fourth leg is a derived character of uloborids and deinopids which is not present in Palaeouloborus. The legs of uloborids bear fine feathery setae amongst the normal setae; neither deinopids nor Palaeouloborus bear them. Palaeouloborus cannot be included in the family Uloboridae because it has neither feathery setae nor fourth tarsal macrosetae. The fossil lacks the many specializations of the deinopids, and in addition the fourth tarsal macrosetal comb is absent. Rather than redefining the family Uloboridae to accommodate the fossil genus, it is left here within the superfamily Deinopoidea, closer to the Uloboridae than the Deinopidae, but not placed in either family. This placement indicates that the loss of a fourth tarsal macrosetal comb is autapomorphic for the fossil genus. Genus PALAEOULOBORUS gen. nov. Derivation of name. Greek, palaios, old, and the living genus, Uloborus, which the fossil genus resembles. Type and only known species. Palaeouloborus lacasae sp. nov.

7 SELDEN: SPANISH CRETACEOUS SPIDERS 263 Diagnosis. Deinopoid with ovate carapace bearing marked break of slope separating anterior of carapace from sloping posterior area; leg 1 more than five times length of carapace and more than twice length of leg 3; many trichobothria on superior?retrolateral surface of femur of leg 2 and superior?prolateral surfaces of femora of legs 3 and 4; paired tarsal claws small, without teeth, median claw long, without teeth, pair of large accessory claws; superior surface of metatarsus of leg 4 in gentle S-shape, proximally convex and then concave, straightening out about half-way along podomere, bearing calamistrum which becomes row of curved bristles towards distal end of podomere; plumose setae present. Row of macrosetae absent from fourth tarsus. Palaeouloborus lacasae sp. nov. Plate 1; Plate 2, figs. 1-5, 7, 8, 10; text-fig. 1 Derivation of name. After Antonio Lacasa-Ruiz, palaeontologist at the Institut d'estudis Ilerdencs, Lérida, Spain. Type specimen. Holotype LP 1755 AP, from the quarry of La Pedrera de Meiá, Sierra de Montsech, north-east Spain, and held in the collections of the Institut d'estudis Ilerdencs, Lérida, Spain. Diagnosis. As for the genus. Description. The carapace shape is determined from fragments of cuticle and from the relief of the matrix. The scraps of cuticle preserved in the carapace region mainly represent coxae and sternum. However, on the left side between legs 2 and 3, cuticle of the carapace edge can be seen lying in a depression caused by pressure of the edge into the matrix. Faint traces of cuticle from the posterior edge of the carapace can also be discerned between the tarsi of legs 3. These cuticle remnants and the faint depression formed by the carapace margin suggests an ovate carapace with a truncated anterior margin. The carapace is 1-73 mm long, and 1-50 mm wide. It is widest just posterior to its midpoint. There does not appear to be a well defined cephalic area, nor an obvious fovea. There is a marked transverse break of slope just posterior to the widest part of the carapace and separating the raised foveal region from the backwardly sloping posterior part of the carapace. This slope is found in some living spiders (e.g. the uloborid Philoponella, PI. 2, fig. 8), and accommodates an abdomen which extends forwards beyond the pedicel. Eyes not seen, no obvious tubercles. The chelicerae are large, 0-67 mm long, and 0-33 mm wide, and somewhat forwardly directed, with parallel sides. The fangs cannot be seen, since the anterior edges of the chelicerae are obscured by overlying matrix. The palps are clothed with setae, and short spines occur on the superior surface of the tarsus. The tarsus is oval, indicating that it was tumid in life. The distal end is not seen, so the presence of a claw cannot be confirmed. The specimen is therefore not a mature male, and could be an immature or a female. Very little of the basal parts of the palp can be seen, but superimposed on the anterior part of the carapace area is a dark line which, under high magnification, is seen to be serrate (PI. 2, fig. 1). This is interpreted as the serrula of the left maxilla; cuticle is absent where the serrula of the right maxilla would have been preserved. The leg formula is The coxae measure approximately 0-58 mm long, and the trochanters 018 mm. Lengths of the more distal podomeres, in mm, are as follows: leg 1: fe 3-27, pa 0-77, ti 1-64, mt 2-50, ta 0-96, total 914; leg 2: fe 211, pa 0-48, ti 1-35, mt 1-44, ta 0-77, total 615; leg 3: fe 1-35, pa 0-48, ti 106, mt 0-48, ta 0-48, total 3-85; leg 4: fe 1-64, pa 0-48, ti 0-87, mt 1-25, ta 0-52, total The legs are clothed with setae of the plumose type. Under high magnification, they are seen to bear a sculpture of striations arranged in a helical pattern, with abundant, minute serrae (PI. 2, fig. 7). No feathery setae can be seen on the legs. The femora are only sparsely setose, setae and bristles becoming more abundant on distal parts of the legs. Groups of trichobothria are certainly present on the femora of legs 2-4, and on the tibia of leg 4. These trichobothria are not feathered (PI. 2, fig. 2). Isolated trichobothria are more difficult to see, and their presence elsewhere on the legs cannot be confirmed. The leg spines are not large, and because only part of the specimen is preserved, the numbers of spines given below are not the maximum number which may be present on the legs. All tarsi are spineless, and bear two small, non-pectinate paired claws, a large median claw which appears to be non-pectinate or if pectinate then with only minute teeth, and two large accessory claws (PI. 2, fig. 3). Fel bears few setae, mainly on the inferior surface (where they are curved) and on the superior surface, especially proximally and distally. Trichobothria may be present on fel (two possible trichobothrial bases can be seen on fel on the right side) but cannot be confirmed. Fel and pal are spineless.

8 TEXT-FIG. 1. Palaeouloborus lacasae gen. et sp. nov., hole-type, LP 1755 AP. Explanatory drawing for PL 1. See Terminology for explanation of abbreviations. EXPLANATION OF PLATE 1 Palaeouloborus lacasae gen. et sp. nov., holotype; Lithographic Limestone, Lower Cretaceous; Sierra de Montsech, Lérida Province, Spain. LP 1755 AP, whole specimen, lower slab, under ethanol, x 21. See text-fig. 1 for explanation.

9 SEiLEJKN, Pti at&uíflfv>rui PIATI; I

10 266 PALAEONTOLOGY, VOLUME 33 Til bears four inferior spines and three superior spines along the shaft of the podomere, one proximal lateral spine, one lateral subdistal spine, and one superior subdistal spine. Mtl bears three inferior and two lateral spines down the shaft of the podomere. Proximally, mtl bears one superior, one lateral, and one inferior spine, and two inferior distal spines are present. Fe2 is spineless and bears few setae, principally on the inferior and superior surfaces. Both second femora are poorly preserved, but a few trichobothrial bases can be seen proximally, apparently on the retrolateral surface. Pa2 is spineless. Ti2 bears bristles proximally and along the inferior surface. Ti2 bears two inferior distal spines, one lateral subdistal spine, one superior subdistal spine, and one superior distal spine. More spines are probably present on ti2 but on both sides of the specimen this podomere is crossed by the large femur of leg 1, which obscures the mid-sections of ti2. Mt2 is very setose, and bears bristles, especially along the inferior surface. Mt2 bears superior and inferior proximal spines, two median inferior spines, and one lateral and two inferior distal spines. Fe3 bears few setae, one tiny spine laterally, and many trichobothria (probably in life) over a large area of the?prolateral surface. Pa3 is spineless. Ti3 is poorly preserved on both sides of the specimen, but bears many bristles. Mt3 bears two short inferior spines and one lateral distal spine. Fe4 is spineless, bears strong, curved bristles distally, and many trichobothria (as many as on fe3) over a large area of the superior,?prolateral surface. Pa4 bears large bristles laterally. Ti4 bears one superior proximal spine, and prolateral, retrolateral and superior spines subdistally. Ti4 bears about four trichobothria on the proximal superior surface, and long, curved bristles distally. The superior surface of mt4 follows a gentle S-shape, proximally convex and then concave, before straightening out about half-way along the podomere. The superior surface of mt4 bears a calamistrum composed of curved setae, apparently in one row, running from the proximal end of the podomere for about two-thirds of its length, where it passes indistinctly into a row of curved bristles which continues to the distal end of the podomere (PI. 2, figs. 5 and 8). Mt4 bears two short inferior median spines, at least two very short spines inferodistally, and large bristles distally. The abdomen measures 3-67 mm long, and 2-89 mm wide. It is ovate, wrinkled posteriorly, and compressed to the right, indicating that it was globose in life. The abdomen is sparsely setose, the setae becoming thicker, but not longer, posteriorly, where they show their plumose structure under high magnification (PI. 2, fig. 7). A pair of subtriangular areas of darker cuticle, each with a small dark patch anteriorly, is situated at the anterior end of the abdomen. These are interpreted as book-lung opercula. Three pairs of spinnerets are visible in the posterior half of the abdomen. They are compressed to the right. Their position indicates that they were not terminal, but ventral, in position in life. A recurved line immediately anterior to the anterior pair of spinnerets represents the cribellum. Little detail can be discerned because the cribellar plate is not preserved (this is presumably on the counterpart), only the fold of cuticle anterior to the plate. (In living spiders the cribellum is commonly invaginated into a fold in front of the spinnerets when not in use, see PI. 2, fig. 6.) Along this fold there are numerous short, blunt setae of a type not seen elsewhere on the spider (PI. 2, fig. 4). Superfamily ARANEOIDEA Latreille, 1806 Remarks. Cretaraneus gen. nov. and Macryphantes gen. nov., described below, are assigned to this superfamily. Both genera lack a calamistrum and cribellum. This does not, by itself, exclude them EXPLANATION OF PLATE 2 Figs. 1-5, 7, 8. Palaeouloborus lacasae gen. et sp. nov., holotype; Lithographic Limestone, Lower Cretaceous; Sierra de Montsech, Lérida Province, Spain, LP 1755 AP, oil immersion. 1, Serrula of left maxilla, bristle at left end, x , Femoral trichobothria, x , Claws on distal end of tarsus of left leg 4; note long median claw and accessory claws, x 100.4, Spinnerets: anterior pair (at top), median pair, and posterior pair (part); recurved line of short setae (seen at left) in front of anterior spinnerets mark approximate position of cribellum; compare with fig. 6, x 85. 5, Proximal part of metatarsus of right leg 4, showing curvature of superior surface with calamistrum; note long, curved bristles at distal end of tibia (bottom left); compare with fig. 10, x60. 7, Plumose seta, x625. 8, Distal end of metatarsus of left leg 4, showing calamistrum (overlying tarsus of right leg 4, on right) terminating in row of curved bristles to left, x 85. Figs. 6, 9, 10. Philoponella sp., Lake Naivasha, Kenya; J. Murphy Coll. No. 1363, under ethanol. 6, Ventral view of posterior end of abdomen of immature male, showing spinnerets and cribellum in front, x 36. 9, Left lateral view of immature male, x , Metatarsus of right leg 4 of mature female, showing curvature of superior surface with calamistrum, x 50.

11 PLATE 2

12 268 PALAEONTOLOGY, VOLUME 33 from a cribellate superfamily because mature males of many living cribellate genera commonly abandon web weaving with the concomitant degeneration of the cribellum and calamistrum, and the preserved specimens of both genera are mature males. However, evidence is presented below that Cretaraneus and Macryphantes show greater resemblance to araneoids than to any cribellate group. Tarsal adaptations (two claws, scopulae, tarsomeres, etc.) for cursorial, saltatorial, and other specialized locomotory habits, as found in the superfamilies Lycosoidea, Clubionoidea, Philodromoidea, Salticoidea, Thomisoidea, and Pholcoidea, are not found in Cretaraneus or Macryphantes, so referral of the fossil genera to any of these superfamilies is rejected. Cretaraneus and Macryphantes also lack the synapomorphies of Scytodoidea (Lehtinen 1986), Dysderoidea (Forster and Platnick 1985), Palpimanoidea (Forster and Platnick 1984; but see Archaeidae, below), and the superfamilies Hersilioidea, Dictynoidea, Hypochiloidea, and Agelenoidea. Until recently, few strong synapomorphies could be mustered to support the monophyly of the superfamily Araneoidea (Coddington 1986). Those characters suggested by Coddington (1986) as araneoid synapomorphies (aggregate and flagelliform glands, paracymbium, serrate hairs, web construction technique) are not only difficult or impossible to see in fossils but also have yet to be thoroughly checked in many extant spiders both within and outside the superfamily. The problem of defining the Araneoidea was highlighted by Millidge (1988), in a discussion of the position of the Linyphiidae. He pointed out that the term 'paracymbium' covers a number of different morphological features on male palps in araneoid and some non-araneoid families, so this character should not be treated as a synapomorphy for the Araneoidea. He also argued that the sticky microdroplets present on the webs of some linyphiids are not produced by the same type of gland (aggregate glands) as in other araneoids, but may be more closely related to the sticky microdroplets found on agelenid webs. Millidge (1988) concluded (on the basis of other evidence as well as that reported here) that the Linyphiidae should be removed from the Araneoidea, and that the superfamily itself may be an unnatural grouping of families which merely share a lack of synapomorphies of other superfamilies. Millidge's arguments have not been accepted by Coddington (1990). Recent work by Coddington (1989a, b, 1990) has amassed a great deal more evidence supporting the monophyly of Araneoidea, including characters of the labium and the spinnerets. Cretaraneus and Macryphantes are assigned to the superfamily Araneoidea. Cladistic analyses of the families within the Araneoidea were attempted by Heimer and Nentwig (1982) and Coddington (1986, 19896, 1990). Shear (19866) included the following major families in the Araneoidea: Theridiidae, Nesticidae, Linyphiidae, Araneidae (including Nephilinae, Metinae and Tetragnathinae), Theridiosomatidae, Symphytognathidae, Mysmenidae, and Anapidae. The latter four have synapomorphies (Forster and Platnick 1977; Platnick and Shadab 1978a, b; Coddington 1986) which are seen in neither Cretaraneus nor Macryphantes, so these families can be discounted. The familial status of Tetragnathidae is discussed below, under that family. The placement of Cretaraneus will be discussed first, followed by that of Macryphantes. The following characters of Cretaraneus suggest the superfamily Araneoidea: broad, pyriform carapace lacking a distinct fovea, presence of a raised cephalic area, globose abdomen, three footclaws with associated serrate bristles, serrate hairs, lack or paucity of trichobothria, paracymbium on the male palp, labium wider than long, and spinnerets in a compact group. Members of the families Theridiidae and Nesticidae possess a comb of serrate setae on the inferior side of the fourth tarsus. Such a feature cannot be seen on Cretaraneus, but since the serrate setae are not always present, or not obviously serrate, in smaller species of living theridiids, the lack of this feature in the fossil genera does not necessarily exclude them from the Theridiidae. Levi and Levi (1962) gave as a diagnostic character for the Theridiidae, chelicerae with up to three teeth on the outer margin and rarely one to three teeth or denticles on the inner margin; Cretaraneus has more cheliceral tee^th than this. One feature of Cretaraneus which suggests a link with the Theridiidae (but not the Nesticidae) is the labium which appears not to be rebordered. Palpal characters in Cretaraneus are not sufficiently distinct to suggest any particular araneoid family; although the simplest palps in the Araneoidea occur in the Theridiidae (Levi 1961), and the

13 SELDEN: SPANISH CRETACEOUS SPIDERS 269 Cretaraneus palp appears to be relatively simple for an araneoid, palps are variable within spider families. Therefore Cretaraneus, though apparently lacking a comb of serrate setae on the fourth tarsus, cannot be excluded with confidence form the Theridiidae. Stridulatory ridges are commonly present on the chelicerae of male linyphiids, but may be obscure in some species, and the labium is usually rebordered. Neither stridulatory ridges nor a rebordered labium is seen in Cretaraneus. A link between Cretaraneus and linyphiids is provided by the raised cephalic area, a feature common in linyphiid males. However, a presumed lateral condyle on the chelicera and accessory tarsal claws (serrate bristles) are features not found in the Linyphiidae, but characteristic of the Araneidae. The rotation of the male palpal sclerites in some Araneidae mentioned by Levi (1983) cannot be confirmed in Cretaraneus because of the rotation produced during fossilization, although the general appearance of the palp in the fossil genus is not reminiscent of the araneid palp. The apparent presence of a median apophysis in the palp of the fossil may provide information on its placement, but at present this character is unresolved (Coddington 19896). The wide labium, presumed cheliceral condyle, and accessory tarsal claws of Cretaraneus are characteristic of most araneoids (Levi 1980). The rather large, forwardly directed chelicerae, and relatively simple palp of Cretaraneus, suggest the metine-tetragnathine lineage, but the fossil lacks other characteristics of this group, and mature males of some linyphiids also possess enlarged chelicerae. Male palps with a superficially simple appearance are characteristic of the genus Nephila (Schult 1983), but Cretaraneus bears few other similarities to the genus. The Jurassic spider Jurataneus rasnitsyni Eskov, 1984 was placed in a monospecific family, the Juraraneidae Eskov, Eskov (1984) defined the family on a unique combination of araneoid characters and could find no apomorphies for the family. Cretaraneus resembles Juraraneus in many ways, including: overall size and shape; leg (but not tarsal) lengths; approximate shapes of sternum, labium, and maxillae; general shape of chelicerae. Cretaraneus differs from Juraraneus in lacking the irregular group of denticles on the inner margin of the chelicera, and the palpal characters interpreted by Eskov (1984) as a large, hook-like paracymbium, a large, pointed median apophysis, and a long, straight conductor (= embolus?). The family Archaeidae is included in the Araneoidea by many arachnologists but, in a radical revision of archaeids and some other small families (for example Mimetidae, previously always placed in Araneoidea), Forster and Platnick (1984) removed them to the Palpimanoidea. They also created the monogeneric families Pararchaeidae and Holarchaeidae for genera previously included in the Archaeidae. Forster and Platnick (1984, p. 99) proposed two synapomorphies for the superfamily Palpimanoidea: cheliceral peg-teeth (modified setae as opposed to cuticular teeth), and an elevated cheliceral gland mound. They also mentioned that peg-teeth are found in some unrelated thomisoid and scytodoid genera, as convergent phenomena, and have been secondarily lost in members of six families assigned by them to the palpimanoids. Cretaraneus possesses true teeth on the chelicerae, but is mentioned here because of some similarities with the Pararchaeidae. The Pararchaeidae differ from the other palpimanoids in having serrate, rather than plumose, hairs, and show similarity with Cretaraneus in the enlarged chelicerae with a prominent keel and large bristles, the pectinate paired foot-claws, uncinate median claw, serrate bristles and lack of an onychium on the tarsus. A cheliceral keel is also found in other groups, for example the Leptonetidae (Gertsch 1974). The male palp of pararchaeids has a strongly developed embolus and a large tegular plate (Forster and Platnick 1984, p. 70), features also apparent in the palp of Cretaraneus. It is possible, therefore, that some relationship exists between Cretaraneus and the Pararchaeidae, which may or may not be an araneoid family. An archaeid spider has been described from the Jurassic (Eskov 1987). Since it seems impossible to refer Cretaraneus to an araneoid family, there are two available options. First, a new, monospecific family could be defined to accommodate Cretaraneus. This course of action would be difficult, given the lack of specialized features displayed by the fossil, and would not provide any additional phylogenetic information. The second option, and the one chosen here, is to leave the genus unplaced within the superfamily Araneoidea. It is possible that future

14 270 PALAEONTOLOGY, VOLUME 33 discoveries will reveal that Juraraneus and Cretaraneus araneoids. Genus CRETARANEUS gen. Derivation of name. Latin creta, chalk, and araneus, a spider. Type and only known species. Cretaraneus vilaltae sp. nov. belong to the same group of Mesozoic Diagnosis. Araneoid spider with subelliptical carapace bearing raised cephalic area and no fovea; subtriangular sternum; small, subtriangular labium; serrate setae covering all parts of body. Chelicerae relatively large (0-4 x length of carapace), forwardly directed (at least in adult male), with inner and outer row of denticles (not peg-teeth), and mesal ridge; male palp with long embolus, and small, proximal?paracymbium; legs relatively equal in length, about three times the length of carapace; femora, tibiae and metatarsi with spines; tarsi with pectinate paired claws, small median claw, and associated serrate bristles; no true trichobothria; globose abdomen. Cretaraneus nov. vilaltae sp. nov. Text-figs Araneae: Lacasa and Martinez, p. 218; pi. 2, fig. 1. TEXT-FIG. 2. Cretaraneus vilaltae gen. et sp. nov., holotype, LC 1150 IEI. Explanatory drawing for text-fig. 3. See Terminology for explanation of abbreviations.

15 SELDEN: SPANISH CRETACEOUS SPIDERS i Owiirirjuw nfmmt Afler Sr Ramon VililU-Olivj L Presidenl of 1he InMitut d'e-studis Herdems, Lefida. 7)f/v ffwi-fawn. Holot>pc- ami only fcnxjwn specimen. I.C F.L complcie specimen on single jwot limestone from quarry of La Cibrui. Sierra de Momscch. north-can Spain: held in colkclwnt of Insiinut d'ewudis lkrdcms. Le'ridfl. Spain, of Dtapwb. As, for ibx genus. DfffrifHitm. A vnhi-pfcvnl'ii -.pij'.-r, nnxl the mulled nf Lhc specimens kni,v,n frimi Monisevh. The carapace cnlwle is pmcnwl lliml IS golden hcoutn in colour. The L-j.rj.pjLC ix IT? mm king -inj 1-^7 mm wide, mil pynform; ils grcjlesl wijlh inxurs ;li I hur-lifllix nf Ihc length behind I he 1111n-. r nurgin I mm ltl.li c-.i vujih TOT-HCi, J. CrVrnmiM'jn Titvfluc ftn. el»p. nm. Ii%>-1l-iL>pt:. Li1li,.igrLiphiL Limestone, IjOwcrCrtlaCCuut.; Stcrrj dc Moncsech. Leridj I'nn-iw. Spain. I.t" N50 IEfI. See icil-fig. 1 far «plan aliom *13.

16 ?.\ L \LONTOLOGY. VOLL M I. S NXI-MI 4. Crtti/ntuw Ilii/Sun- UEN VI -.P NOV.i HOLOJYR* L.IIHNAR.IPLIBC' UMEVICMC. LOWER CRCIANOTA;SON & MONTWH. LERIDA PROVINCE. SPAIN. LC 1150 IEI. *. ORSUL CLNW* ON fcfl LEG 2:RAHELARGE, PCCIIRUTC. PAIRED %-LAW\. SNULL, UNTITLE, MEDIAN DAW (BOTTOM CENTRE). AND CURVFLJ. SERRATE BRALLES. ESPECIALLY INTERIORLY. «220. LYPICAJ SPUE, ON LIBU OF KFT KG I. ITWWINI HIHIKH *440. C, RIAHL PULPUL LIBIA AND TIIRWS; SEE JT FOR EXPLAWHKBN. J, KFT PALPAL TARSUI; SEE/FOR CAPLNNATION. xsb. R. LAML CLAW; ONRIGHT LEG. 4. x 220. /. CIPLJN;II DRAW INP OF J.T\ CXNLJN;HNA-> DRAWING OF*. SEC RRNIOUFAGP FOR ESPLANALTON OF ABBRCVIALTONS. I TIE CARAPACE EDGES EORN-ERAC IN VERY GEHK WNTW(UNWLIJHIE ROUNDED ANLERINR NXIRJIM. AR>D IN -AN'-IBR SHAPED CURVES WHICH CONVERGE MORERAPIDLY 1C THE POSTERIOR MARGIN. THE POSTERIOR MAROIN h BROAD AND HAS A DRHINCI MEDIAN ERAB-IS NU.*M. THERE I* NO PROMINENT fa VEAL JEPNJISION, FR*NN»HKH II CAN he CONCLUDED THAI I FOVEA MS PRNH.IHK AH-*,-NL ABOUT O:N.'-L UA.RTCRORTHC WAY hack FRUM I HE ANTERIOR MARGIN OR TL*E CARAPACE, ON EACH LATERAL M.Lrgin. IS SLIGHT EMHAYMENT AND a DARK PALCH OF LTJIL-TIC WHICH FORM\ A LINEAR FENLURE EXTENDING RADIALLY

17 SELDEN: SPANISH CRETACEOUS SPIDERS 273 inwards. This is interpreted as a shallow sulcus which reflects a condyle for articulation of the chelicera on the inner surface of the carapace. A radial crack in the posterior right-hand quadrant indicates some crushing of the carapace. Dark areas in the anterior half form a V-shape pointing forwards. The V-shape is asymmetrical with respect to the midline, the left limb being more parallel to the midline than the right, and the whole shape is left of the midline. Some folding is associated with the limbs of the V-shape. This shape is interpreted as a sulcus or break in slope separating the peripheral parts of the carapace from a raised cephalic area in the anterior half of the carapace. The asymmetry of the V-shape, in contrast to the symmetry of the rest of the carapace, indicates left-lateral compression of the cephalic area during compaction of the sediment. A small, semicircular, dark area of cuticle approximately centrally placed near the anterior border of the carapace is interpreted as the posterior border of a right eye, the anterior half of which is missing. Other eyes, if similar in size, are most likely to be obscured by the dark lines of crumpled cuticle around the anterior edge of the cephalic lobe. The whole carapace is separated from the ventral prosoma and moved slightly to the left, exposing the right coxae. A pair of dark lines in the posterior half of the carapace, subparallel to the midline but diverging towards the anterior, are interpreted as left and right edges of the sternum. The anterior border of the sternum is seen as a transverse, recurved line just anterior to the midpoint of the carapace. Due to the left-lateral movement of the carapace, the sternum appears mainly to the right of the midline. The sternum is widest anteriorly, and narrows gradually to a blunt point situated between the coxae of the fourth pair of legs. In front of the anterior margin of the sternum are some rounded dark areas. The somewhat triangular dark area anterior to the midline of the sternum is interpreted as the labium. The labium is widest posteriorly, where it is distinctly separate from the sternum. The lateral edges converge to a rounded anterior margin. The two areas to the right of the sternum represent the right maxilla with the anterolateral carapace sulcus superimposed. The posterior part of the left maxilla can be seen to the left of the labium, but its anterior part is obscured by the right limb of the V-shape surrounding the carapace cephalic area. No serrulae can be seen on the maxillae; there are many setae visible in the intermaxillary area. The chelicerae are about 0-69 mm long, and project forwards in front of the carapace. Their lateral sides are straight and parallel to each other, their inner sides are convex and partly overlap in the fossil. Their dorsal (superior) surfaces bear numerous short, stiff setae; laterally, curved bristles are present. Each chelicera bears two rows of denticles extending from the mesal side of the anterior border to about half-way down the inner edge, the outermost row bearing at least three and probably five denticles, the inner row with at least two denticles. The fangs are not preserved, and were presumably on the counterpart which was not collected. A prominent ridge, or keel, runs from the end of the tooth row (which is ralatively short, about one quarter the length of the chelicera) along the mesal surface. No stridulatory ridges can be seen. The presence of a thickened sulcus on the anterolateral side of the carapace (see above), which probably reflects an internal apophysis for articulation of the chelicera, suggests the presence of a condyle on the chelicera; the condyle itself is not preserved on the specimen. From the morphology of the preserved cheliceral parts, it is apparent that the fangs worked transversely (labidognathous). This specimen is a mature male because the palps are modified for the transmission of sperm (text-fig. 4c, d,f, g). Both palps are bent over to the right due to the left oblique compression of the specimen. The tarsus of the left palp now appears to the right of the right chelicera, and the right palpal tarsus lies beneath the femur of right leg 1. The appearances of the sclerites on each palp differ because the palps are compressed in different ways. The left palp presents a mesal view, and the right an ectal view. The total length of the palp, from the maxilla to the tip of the bulb (i.e. excluding the embolus, see below) is approximately 2-20 mm. The palpal femur is about equal in length to the adjacent chelicera. The patella and tibia are covered with long setae. The tibia is a distally expanded, triangular podomere, bearing long bristles which radiate distally to partly cover the tarsus. Distal to the tibia is an ovoid body with numerous sclerites superimposed on it. The ovoid body is interpreted as the superimposed bulb and cymbium (modified tarsus). The cymbium is not separately recognizable from the bulb, and is therefore presumed to be no longer than the bulb and related parts. The right palp seems to present an approximately ectal view, and the left palp an approximately superomesal view. Immediately distal to the tibia, an elliptical dark area may represent a small, separate paracymbium. On the right bulb, a curved, lath-like sclerite extends from the superoproximal edge to the inferior side of the bulb. This may be the tegulum. On the left palp, the different direction of compression has caused this structure to appear curving from the apparent inferior edge towards the superodistal direction. On the right palp below the supposed tegulum is a rather complex, hooked structure, also visible on the left palp. This may represent a median apophysis. The interpretations of both tegulum and median apophysis are uncertain. The gently helical, acuminate structure, as long as the bulb itself, extending distally from the distal end of the bulb is interpreted as the embolus. It is easily seen on the left palp, but on the right palp only its basal part is visible, the remainder

18 274 PALAEONTOLOGY, VOLUME 33 disappearing into the matrix beneath. It is possible that this structure represents the conductor (see, for example, the helical conductors of some tetragnathines, Levi 1981) in which case the embolus cannot be seen. The legs are short and nearly equal in length. The leg length formula is The coxae are visible on the right side. Each is about 0-39 mm long, and slightly longer than broad. The trochanters are very short (about half as long as broad), and each is about 0-19 mm in length. Approximate lengths of more distal podomeres and total leg lengths, in mm, are as follows. Leg 1: fe 1-73, pa 0-46, ti 1-64, mt 1-08, ta 0-58; total Leg 2: fe 1-54, pa 0-46, ti 1-50, mt 117, ta 0-69; total Leg 3: fe 1-25, pa 0-46, ti 104, mt 0-87, ta 0-52; total Leg 4: fe 1-31, pa 0-46, ti 1-25, mt 0-92, ta 0-52; total All legs are thickly clothed in long, fine setae, and the femora, tibiae, and metatarsi bear spines. The setae are not plumose, under high magnification, but some appear to be sparsely serrate, and on the larger ones a rectilinear pattern, as observed on the spines, can be seen. Among the leg setae, especially on the distal podomeres, there are a few short and thin, gently S-shaped, rather crumpled, setae which extend outwards from the podomeres at high angles, and are set in follicles which appear rather too large for the thickness of the seta. Some of these hairs are interpreted as chemosensory setae, as described by Foelix (19706). Others may have had a function akin to that of trichobothria; no undoubted trichobothria can be seen. The spines have a rectilinear sculpture (text-fig. 4b). The numbers of leg spines given here are the maximum number which can be seen on the specimen. This must be considered to be either the actual or the minimum number present in life, since in this specimen the counterpart was not collected, and some spines may be preserved on the counterpart only. However, since by far the greater part of both dorsal and ventral sides of the animal is preserved on the slab, it is considered likely that few spines are unaccounted for. The femora bear stout, curved spines and some curved bristles in the distal area. Four spines are present on fel, all apparently on the superior side of the podomere, of which two are subdistal and two distal in position; there are three on fe2 and fe3, superior in position; there is at least one on the posterior side of fe4. No other spines are present on the femora. The patellae are spineless. The tibiae and metatarsi bear spines in three areas: proximal, median, and distal. The distal spines are stout, curved, and inferoanterior and inferoposterior in position; the others are longer, straighter, and occur on all sides of the podomeres. In addition, stiff bristles occur in the median and distal areas of these podomeres. Til bears five spines in a whorl on the proximal area, four (supero- and inferoanterior and supero- and inferoposterior) in the median area, and two distally. Ti2 bears four spines (one superior and three inferior) in the proximal area, two superior median, and two in the distal area. Ti3 has no proximal spines, five in a whorl in the median area, and two distally. Ti4 bears at least five spines in various positions in the proximal and median areas, and two distal spines. Mtl bars five proximal spines, two pairs in the median area, and two distally. Mt2 has four proximal spines, two in the median area, and two distally. Mt3 has three spines in each of the three areas. Mt4 bears three proximal spines, probably four in the median area, and at least three distal spines (one superior in addition to the usual anterior and posterior). All tarsi are spineless. Distally, pectinate paired claws are present, each with one row of nine teeth, and a small, unciform median claw, but no onychium (text-fig. 4a, e). Surrounding the claws are a number of serrate bristles; these are gently S-shaped at the end with the serrations on one side. There is no comb of serrate bristles along the shaft of this podomere. The abdomen is 2-17 mm long and elliptical to subcircular in outline. Very little cuticle is preserved, so it is presumed to have been thin in life, but the abdomen outline is clearly seen due to its covering of setae. Cuticle between the setae can only be seen in ill-defined areas mainly in the left half of the abdomen; these are interpreted as?dorsal pigmented areas in life. Generally darker coloration and greater density of setae indicate position of spinnerets which, however, are not separately discernible. The spinneret region can be seen on the right side of the abdomen (due to the left oblique compression) and appears to have been subterminal in position. A longitudinal dark streak left of the midline between the carapace and abdomen is presumed to represent the remains of the pedicel. Family TETRAGNATHiDAE Menge, 1866 The familial status of Tetragnathidae has been supported by many authors (for example, Kaston 1948, 1978; Locket and Millidge 1953), but resisted by others (Levi 1980; Roberts 1985, p. 198) in the past. The situation is further complicated by the position of the Metinae, placed by some in Tetragnathidae and by others in Araneidae. The most recent opinions of Levi (1986), Coddington (19896), and Platnick (1989) are that tetragnathines, metines, and nephilines should be placed together in the family Tetragnathidae, separate from the Araneidae. In the past, these subfamilies